Textured Enhanced Image De-noising using Fast Wavelet Transform

Abstract

The digital imaging technologies possess very hasty development that uses Giga-pixels to store an image. Image de-noising algorithms plays a significant role in the restoration process. In an image the texture regions are homogeneous and are composed of local descriptor, a trade off exist between visual quality of image and the enhanced texture regions. In the existing paper the Gradient Histogram Preservation (GHP) method on the enhanced image regions have a limitation where it cannot be directly applied to nonadditive noise removal such as Multiplicative Poisson Noise (MPN) and Signal-Dependent Noise (SDN) to overcome the limitation Fast Wavelet Transform is used. In this work an image is first added with different types of noise like Additive White Gaussian Noise (AWGN), Salt and Pepper Noise, Poisson Noise, Signal Dependent noise and Flicker Noise, the noisy image is restored using filters, next the enhanced texture region of the image is chosen which is blurred or deformed and the fine details of the texture is obtained using Fast Wavelet Transform (FWT). The proposed work is analyzed in Frequency domain by considering various parameters like Peak Signal to Noise Ratio (PSNR), Correlation Factor (CF) and Standard Deviation (SD) and the quality of the enhanced region of the image is improved to the best level than the conventional noise removal algorithms.